Using link types in Web page ranking and filtering

Corresponding to the evolution of the Web from the poorly structured towards more structured, semantic rich network, search methods that apply to it are also evolving from using little structural and semantic information to using more such information that is available. This paper proposes a ranking and filtering mechanism that makes use of link types that is representable with new Web standards. We suggest that page ranking can be propagational through links and the propagational rates depend on the types of the links and users' specific set of interests. Page filtering can be decided based on link types combined with some other information relevant to links. For either a ranking or filtering task, a profile containing a set of ranking or filtering rules to be followed in the task can be specified to reflect users' specific interests. Technical issues in implementing the mechanism in a search system are also discussed.